Quick Start¶
Get DAM running without robot hardware, then confirm that the console, CLI, and demo Stackfiles are wired correctly. Budget about 10 minutes on a fresh machine; repeat runs are faster.
What Success Looks Like¶
By the end of this page you should have:
- A local Python environment and Rust extension built by
make setup - The backend API listening on
http://localhost:8080 - The DAM Console available at
http://localhost:3000 - Example Stackfiles passing validation
- A clear next page for learning guards, Stackfiles, or the console
Installation & Setup¶
# Clone the repository
git clone https://github.com/ez945y/DAM.git
cd DAM
# One-time setup (Python venv + Rust extension + npm)
make setup
Expected result: setup finishes without errors and creates .venv/. It may take several minutes because it installs Python packages, builds the Rust extension, and installs console dependencies.
If setup fails because cargo is missing, install Rust from rustup.rs and run make setup again.
Start DAM¶
This launches:
- API server at http://localhost:8080
- DAM Console at http://localhost:3000
- Runtime host using the repo's default Stackfile convention
Confirm It Is Running¶
Open your browser:
- Dashboard: http://localhost:3000
- API Docs: http://localhost:8080/docs
Expected result: the dashboard loads and shows runtime status, cycle counters, risk state, guard decisions, and latency panels. If the console opens but looks disconnected, confirm that the API server is still running on port 8080.
Validate A Demo Stack¶
Before editing a robot configuration, validate the example Stackfiles:
Expected result: every file under examples/stackfiles/ reports OK.
Inspect one demo Stackfile to see what DAM resolved:
Expected result: the output lists safety settings, guards, boundaries, tasks, and fallback escalation. You do not need to understand every field yet; just confirm that the stack can be loaded and explained.
No hardware path
The demo Stackfile is safe to inspect and validate without a robot. It uses a dataset source and a matching sink reference for replay-style development. The ACT policy may use CPU and may download model assets the first time it is run.
Learn The First Workflow¶
Use this loop when you are learning DAM:
- Pick an example Stackfile under
examples/stackfiles/. - Run
make validate. - Inspect it with
.venv/bin/dam inspect <path>. - Start DAM with
make run. - Open the console and identify whether the latest cycles are PASS, CLAMP, or REJECT.
The important habit is to validate configuration before thinking about hardware.
Optional: Run Tests¶
# Full test suite (Python + Rust + Frontend)
make test
# Or test individually
make test-py # Python only
make test-rs # Rust only
make test-ui # Frontend only
make lint # Linters only
Customize with Stackfiles¶
After the demo path works, edit a copy of an example Stackfile to change:
- Guard parameters (L0–L3)
- Task boundaries
- Hardware adapters
- Safety constraints
Then validate before running:
Available Commands¶
| Command | Purpose |
|---|---|
make setup |
First-time setup (Python + Rust + npm) |
make setup-lerobot |
Setup with SO-ARM101 hardware support |
make run |
Start backend + frontend |
make test |
Run all tests + linters |
make clean |
Remove build artifacts |
Next Steps¶
- Learn DAM -- guided learning path
- Safe Recording -- guard actions during IL data collection
- Stackfile Walkthrough -- read a Stackfile field by field
- Use Cases -- how DAM helps at each stage of the ML pipeline
- Troubleshooting -- fix first-run issues